Vasileios Belagiannis

• Always-on Deep Neural Networks

  (Third Party Funds Single)

  Term: 1. March 2023 - 28. February 2026
  Funding source: DFG-Einzelförderung / Sachbeihilfe (EIN-SBH)
  Abstract

  Computer vision contributes in creating visual priors as self-contained tasks or input to another system. In the context of autonomous navigation, the system can be a mobile agent that not only relies on the raw sensory inputs, but also on computer vision algorithms for understanding the environment. Recent studies on embodied agents show that an agent acts more accurately when visual priors such as semantic segmentation, depth estimation are provided next to the raw input data. Producing the visual priors though comes at the cost of data collection and annotation. The latest approaches build on deep neural networks, which are trained with supervision. For that propose, a large pool of data and annotations has to be created prior to training the model. To address this limitation, simulation is an alternative source for data and annotation generation. In the context of deep neural networks, it can be considered for the replacing the real-world, where a large amount of synthetic data is created according to the task in place. Although, the data simulation has clear advantages over the real-world datasets, there is also a clear limitation. Training a deep neural network with synthetic data does not result in good performance on real-world data.In this research project, we are going to conduct research on closing the performance drop when transferring deep neural network models from the simulation to real-world applications. Our testbed for measuring the performance will be semantic image segmentation and depth estimation from a single image. In our research, we will propose algorithms that teach a deep neural network how to fast learn adapting into new environments. This concept is widely known as meta-learning. In this project, it will be explored for learning a model in simulation and then transferring it to the real-world. Meta-learning has never been seen as a way to tackle model transfer, but its formulation suits well to the problem.
  →More information

• Transfer von tiefen neuronalen Netzen von der Simulation in die reale Welt

  (Third Party Funds Single)

  Term: 1. December 2022 - 30. November 2025
  Funding source: DFG-Einzelförderung / Sachbeihilfe (EIN-SBH)
  Abstract

  Computer vision contributes in creating visual priors as self-contained tasks or input to another system. In the context of autonomous navigation, the system can be a mobile agent that not only relies on the raw sensory inputs, but also on computer vision algorithms for understanding the environment. Recent studies on embodied agents show that an agent acts more accurately when visual priors such as semantic segmentation, depth estimation are provided next to the raw input data. Producing the visual priors though comes at the cost of data collection and annotation. The latest approaches build on deep neural networks, which are trained with supervision. For that propose, a large pool of data and annotations has to be created prior to training the model. To address this limitation, simulation is an alternative source for data and annotation generation. In the context of deep neural networks, it can be considered for the replacing the real-world, where a large amount of synthetic data is created according to the task in place. Although, the data simulation has clear advantages over the real-world datasets, there is also a clear limitation. Training a deep neural network with synthetic data does not result in good performance on real-world data.In this research project, we are going to conduct research on closing the performance drop when transferring deep neural network models from the simulation to real-world applications. Our testbed for measuring the performance will be semantic image segmentation and depth estimation from a single image. In our research, we will propose algorithms that teach a deep neural network how to fast learn adapting into new environments. This concept is widely known as meta-learning. In this project, it will be explored for learning a model in simulation and then transferring it to the real-world. Meta-learning has never been seen as a way to tackle model transfer, but its formulation suits well to the problem.

  →More information

Lectures

• Machine Learning in Signal Processing
  • Further information is available on StudOn and Campo.
• Introduction to Deep Learning
  • Further information is available on StudOn and Campo.
• Advanced Topics in Deep Learning
  • Further information is available on StudOn and Campo.
• Perception in Robotics
  • Further information is available on StudOn and Campo.

Guide to scientific work

• Guide to scientific work
  • Further information is available on StudOn and Campo.

Seminars

• Seminar on Selected Topics in Multimedia Communications and Signal Processing
  • Further information is available on StudOn and Campo.
• Seminar on Selected Topics in Machine Learning
  • Further information is available on StudOn and Campo.
• Seminar über Bachelor- und Masterarbeiten
  • Further information is available on StudOn and Campo.

Lab Courses

• Machine Learning in Signal Processing
  • Further information is available on StudOn and Campo.